Note: Two updates added shortly after publication.
You don’t need Wednesday’s spreadsheet for the newest Minnesota Public Radio/Humphey Institute poll: Mark Dayton has a statistically significant 12-point lead over Tom Emmer, 41-29 percent, with Tom Horner sinking to 11 percent.
The 751-likely-voter, four-day survey has a margin of sampling error, including survey design effects, of 5.5 points plus or minus. It does not include “cell phone only” voters who lack land lines.
Downbeat Republicans who look at Emmer’s support will want to focus on a few things.
The Humphrey Institute is not only tough on itself when reporting sampling error, it lists a relatively high percentage of undecideds/refused-to-responds: 20 percent. That’s a big pool for a trailing candidate to draw from.
Update: HHH pollster Larry Jacobs says if they allocate leaners, Dayton’s margin rises to 16 points, with 9 percent undecided. I haven’t heard anyone mention that kind of margin until now.
Then there’s the partisan breakdown: 45 percent Democrats, 39 percent Republicans, 16 percent independents. With Republicans motivated throughout the country, any poll with that many more likely Democrats voters is unusual. However, the Humphrey Institute aggressively allocates “leaners”: independents must resist party preferences in two separate questions to stay that way.
Finally, the Humphrey Institute’s poll is one of many that has rated DFL gubernatorial candidates ahead in the final week, only to see them lose on Election Day. The final 2006 MPR/HHH gubernatorial poll had DFLer Mike Hatch up by 6, the biggest margin of the final major polls; he lost by 1. Obviously, Dayton’s final-poll cushion is double Hatch’s.
(Another college-based poll, St. Cloud State’s, was a bigger 2006 outlier; it had Hatch +10 but was conducted a bit earlier. This year’s version also has Dayton up by 10.)
Update: Unlike other ’06 final polls, MPR’s was taken before Hatch’s infamous “Republican whore” blowup. However, no other major pollster (SurveyUSA, Rasmussen, Mason-Dixon) had Hatch up by more than 2 in the two weeks before WhoreGate.
The MPR poll was well within the error margin on the 2006 and 2008 U.S. Senate races. However, in the 2008 Presidential poll, its 19-point Obama lead was higher than his ultimate 10-point win.
In every race, the MPR poll gave Democrats a higher total than they received on Election Day. However, “final” polls, especially ones that survey a week before most people vote, cannot always capture late deciders.
Currently, Dayton’s leads in the other major surveys are: 3 point (Rasmussen), 5 points (SurveyUSA/KSTP) and 7 points (Star Tribune). KSTP will release a new SurveyUSA poll on the governor’s race tonight.
Gee, Dave; despite the obvious opportunities for comedic relief, a Dayton win would be a disaster for me and my family…why am I not concerned?
“The poll’s sample includes 45 percent Democrats, 38 percent Republicans and 16 percent independents. The percentages for both the Democrats and the Republicans are higher than recent Star Tribune Minnesota Polls, which had a sample that included roughly a third of voters in each category.”
Oh yeah, that’s right.
It would seem wise to either pay no attention to the left-leaning MPR/HHH polls, or routinely subtract 5 to 10 points from the DFL candidate’s total.
In this case, subtract 12 points and you will be closer than half a percent the morning after election day. You called in John.